Jérémi Régnier

Kalman-Filter-Based Indicator for Online Interturn Short Circuits Detection in Permanent-Magnet Synchronous Generators

Kalman filtering has been largely used in adaptive control for motor drive applications. However, few works relate to applications in the field of monitoring. In this paper, a fault indicator for interturn short circuits detection in permanent-magnet synchronous generators (PMSGs) is introduced. Based on the extended form of the Kalman filter and associated to an appropriate model of PMSG expressed in the

On-line parameter estimation of PMSM in open loop and closed loop

dq

Stator current based indicators for bearing fault detection in synchronous machine by statistical frequency selection

frame, this indicator has been developed through the identification of parameters specifically introduced to point out stator faults. Special attention has been paid to the indicator sensitivity and its robustness to the generator operation mode, which implies a large range of operating points. The indicator performances have been first evaluated using an appropriate simulation model. A complete experimental campaign has been performed to validate previous simulation results. A faulty PMSG, specifically built for stator fault experiments, has been associated to a digital signal processor for an online implementation of the proposed indicator. Finally, the issue of parameter uncertainties is solved by introducing a variable threshold for detection improvement.

Monitoring of turn short-circuit faults in stator of PMSM in closed loop by on-line parameter estimation

The paper discusses the multi-models approach for on-line parameter estimation of permanent magnet synchronous motor (PMSM). The study is focused on a PMSM with non sinusoidal electromotive force (emf) in open loop and closed loop conditions. The electrical parameters of 2-phases models in Extended Park frame are estimated by using the recursive least squares algorithm (RLS).

On-line inter-turn short-circuit detection in permanent magnet synchronous generators

The aim of this paper is to present some indicators developed for efficient detection of bearing defaults in high speed synchronous machines using a stator current analysis. These actuators are used in an air conditioning fan for aeronautic applications. The signatures of the bearing defects appear through an increase in amplitude of specific current harmonics multiples of the rotation frequency. From an experimental comparison between a healthy fan and another with damaged bearings, an automatic frequency selection is performed to identify the frequency ranges for which the energy is the most sensitive to the considered faults. From these frequencies, several strategies are investigated to propose a suitable indicator for the bearing fault detection. A post-processing algorithm is then developed and tested for different measurements, different types of faults and different operating points, to ensure the robustness of the proposed method.

Wigner distribution for the diagnosis of high frequency amplitude and phase modulations on stator currents of induction machine

This paper presents a method for monitoring of turn short-circuits faults of permanent magnet synchronous motor (PMSM) with non sinusoidal electromotive force (emf) in closed loop conditions. The proposed method is formulated by combining multi-model approach in Extended Park frame and the recursive least squares algorithm (RLS). The fault and its severity are indicated by the surface of structure distance between the Extended Park model and PMSM.

Experimental comparison between diagnostic indicators for bearing fault detection in synchronous machine by spectral Kurtosis and energy analysis

This paper focus on inter-turn short-circuit detection in Permanent Magnet Synchronous Generators (PMSG). Inter-turn short-circuit current are among the most critical in the PMSG. For safety considerations, a fast detection is required when a fault occurs. In this work, a fault indicator for on-line diagnosis is proposed. Based on a faulty PMSG model expressed in Parks reference frame, the number of short-circuited turns is estimated using Extended Kalman Filter (EKF). Simulation and experimental results are provided to demonstrate the quickness and the robustness of the fault indicator with regards to various operating points on an electrical network.

Bearing fault detection in synchronous machine based on the statistical analysis of stator current

This paper deals with mechanical fault monitoring in induction machines from stator current measurements. The considered faults lead to amplitude and/or phase modulations of the measured stator current. The different faults can be characterized by their time-frequency signatures via the Wigner distribution. The time-frequency representations apply to complex signals that may be obtained through the Hubert transform of the real measured signal. In case of high frequency modulations, it can not be taken advantage of the time frequency signatures. This study proposes an alternative complex signal representation for modulated stator currents. The so-called space vector is obtained through the Concordia transform. From three stator current measurements, the Concordia transform builds a complex vector which conveniently carries the information about phase and amplitude modulations. This paper applies and compares the Wigner distribution computed with the Hilbert and Concordia transforms in case of simulated and experimental signals with various modulation frequency ranges.

A generic diagnosis protocol for the monitoring of induction motors based on multiple statistical references in the torque-speed plane

In this paper, some indicators are developed for efficient detection of bearing defaults in high speed synchronous machines. These indicators are based on the analysis of stator current. As bearing defect signatures can be tracked through amplitude increase of some current harmonics, two specific indicators have been built based on energy considerations and on the Spectral Kurtosis analysis. These indicators are tested on a real industrial fan equipped with ceramic balls, in its environment. Several measurements for different operating points are tested to validate the approach and to its robustness during long time tests. From an experimental comparison between a healthy fan and another with damaged bearings, a frequency selection is performed to identify the frequency ranges where the energy is the most sensitive to the considered faults. This actuator is used in an air conditioning fan in aeronautic applications.

A Fast Approximation Algorithm for Set-Membership System Identification

In this paper, an original method to compute an indicator for efficient detection of bearing defaults in high speed synchronous machines is presented. This indicator is based on the statistical analysis of the stator current spectrum. The principle of the method is to compare the current spectrum to a healthy reference spectrum. The reference spectrum is used to center and reduce the current spectrum and provides an indication on its difference from the reference spectrum. A statistical based indicator is then constructed as the sum of the contributions of the centered reduced spectrum for different interesting frequency bands. This indicator has been tested on 2 different test campaigns, for 4 different speeds and compared a vibratory indicator. Results show that the proposed indicator evolves the same way than the vibratory indicator and provides with an efficient detection of bearing fault with only very few false alarms.